# 10 AI Services Businesses Pay For: The Full Playbook

A free resource from Mariah Brunner, learnaiwithmariah.com/guides/smb-services/

Pricing below is suggested starting pricing for productized services, not a guarantee of income. Your results depend on your niche, your network and how many people you pitch. Replace every bracket and every result in a pitch with your own real numbers.

## The one rule
Pick ONE service. Sell the same thing to 10 clients in a row. Add a second service only after the first has 3 to 5 paying customers.

## Your 30 / 60 / 90 day path (same for every service)
Days 1-30: pick a narrow niche. Build a free or discounted version for one friendly business in exchange for a testimonial and a case study. Track real before and after numbers (hours saved, tickets handled, reply rate). Record a 5-minute Loom of it working.
Days 30-60: send 20 to 50 personal DMs a week to the buyer below, leading with the case study. Land paid client #1. Deliver, document, ask for a referral.
Days 60-90: turn it into a fixed package with a name, a price, a timeline and locked deliverables (for example "Support AI Sprint: $5K, 2 weeks"). Offer every client a monthly retainer for updates.

## Platform notes (checked November 2026)
- Sharing a Claude Project with a team needs a Claude Team or Enterprise plan. On Pro and Max you can only share chat snapshots.
- Nothing you build on Claude's cloud keeps data "inside the client's environment." For sensitive clients (legal, HR, health), put them on Claude Team or Enterprise under commercial terms, and get their IT or security sign-off first.
- Connectors exist for Gmail, Google Calendar, Google Drive, Slack and many CRMs. Check the client's tools against Claude's connector directory before you promise an integration.

---

## 1. Custom Claude assistant for HR teams
Suggested price: $8-15K build, $1-3K/mo retainer, 2-3 weeks.
Who pays: HR directors at 50-500 person companies, people ops leads at fast-growing startups, office managers who own HR by default, outsourced HR firms.
Deliver: discovery interview plus the handbook, policies, benefits docs and the top 50 questions with current answers. A Claude Project (or skill) with all docs loaded and instructions for tone, escalation and when not to answer. A 30+ question test bank. A rollout guide (team sharing needs Claude Team or Enterprise). 30 days of tuning, then quarterly refreshes.
Pitch: "Hi [Name], I build AI assistants for HR teams that answer the most common employee questions (PTO, benefits, policy) from your own handbook. Can I send a 5-minute Loom of how it works for a [company size] team?"
Pitfalls: position it as handling tier-1 questions so HR can focus on hard ones. Always escalate harassment, discrimination, terminations and anything legal to a human.

## 2. Sales email drafting system
Suggested price: $5K setup, $2K/mo retainer (or about $1K per rep per month for teams), 2 weeks.
Who pays: sales managers at 20-200 person SaaS companies, founder-led B2B startups doing outbound, sales agencies.
Deliver: a prospect research workflow (company name or URL in, research brief out). 5-7 email templates by use case. A personalization Project that combines research, template and the rep's own past emails. 3-5 step follow-up logic. A simple tracking sheet for opens, replies and meetings. A 2-hour training.
Pitch: "Hi [Name], I build AI outreach systems that research each prospect and draft emails in your reps' own voice. For [case study client] it [your real result]. Want a 5-minute Loom?"
Pitfalls: train on each rep's real voice. Keep a human review before every send. No spammy volume plays.

## 3. Internal AI training workshops
Suggested price: $3-5K half day, $5-10K full day, $10-20K for a multi-session series, 2 weeks prep.
Who pays: HR and L&D directors, COOs, department heads, mid-size law, accounting, agency and real estate firms.
Deliver: a pre-workshop survey. A custom deck with role-specific examples. Live demos and 3-5 hands-on exercises. A take-home sheet of 10-15 prompts for their industry. The recording. 30 days of follow-up questions. Optional office hours 30 days later.
Pitch: "Hi [Name], I run hands-on AI workshops for [industry] teams. Half a day, built around what your team actually does, with live demos and practice. Can I send the syllabus?"
Pitfalls: niche the title ("AI for Sales Teams: 90 Minutes That Save Hours Every Week" sells, "AI Workshop" doesn't). Live demos beat slides. Always pitch a follow-up.

## 4. Sales prospect research briefs
Suggested price: $5-8K setup, $1.5-3K/mo retainer, 1-2 weeks.
Who pays: sales VPs at B2B SaaS, heads of business development at agencies, account-based teams, founders doing their own sales calls.
Deliver: a workflow that turns a calendar event or prospect name into a 1-page brief (who they are, last 60 days of news, priorities, likely pains, decision makers, one smart question, 3 likely objections), delivered before each call by email or Slack. Calendar and CRM connectors where available. A 1-hour training. A quarterly review.
Pitch: "Hi [Name], I build automatic pre-call briefs. Your reps walk into every meeting with a 1-pager on the prospect. Want a sample brief?"
Pitfalls: every brief must cite recent, specific sources, and say "nothing found" instead of guessing. Don't scrape LinkedIn. Use public sources and data the client already licenses.

## 5. Customer support AI for Shopify stores
Suggested price: $5K setup, $500-1.5K/mo retainer, 2 weeks.
Who pays: Shopify owners doing roughly $500K-$10M a year without a support team, DTC brands hiring their first support person, Shopify agencies.
Deliver: a Claude Project loaded with the catalog, shipping and return policies, FAQs, brand voice samples and the top 50 past tickets with ideal replies. Escalation rules. Drafted replies inside their help desk or chat tool (Gorgias, Tidio, ManyChat and similar). Tracking for tickets handled, response time and satisfaction. Monthly updates.
Pitch: "Hi [Name], love [product]. How much of your week goes to support tickets? I set up AI that drafts answers to your most common questions in your brand voice. Want a 5-minute Loom?"
Pitfalls: refunds, complaints and angry customers always go to a human. Measure the client's real deflection rate instead of promising one.

## 6. First-pass contract review for law firms
Suggested price: $10-20K setup, $2-5K/mo retainer, 3-4 weeks.
Who pays: boutique firms (business, real estate, employment law), 10-50 attorney firms, small in-house legal teams.
Deliver: a Project built on the firm's templates and review checklist. A structured first-pass review of key terms, deadlines, renewals, liability caps and unusual clauses. Deviation flags against the firm's standard. A draft-vs-template comparison. A plain-English summary for clients. Training for paralegals and associates. Monthly tuning.
Pitch: "Hi [Name], I build first-pass contract review assistants for boutique firms, set up on your own templates. Your attorneys keep final judgment, the AI does the first read. Can I show you a 5-minute demo on a [contract type]?"
Pitfalls: never say it replaces lawyers. Data handling is the whole sale: use Claude Team or Enterprise, document your security stance, and get the firm's sign-off. Specialize in 1-2 contract types. An attorney reviews every output.

## 7. Onboarding assistant for new hires
Suggested price: $8-15K setup, $1-2K/mo retainer (or a per-hire fee for high-volume hiring), 2-3 weeks.
Who pays: people ops at 20-200 person companies hiring every month, VPs of operations, engineering leads, high-turnover businesses.
Deliver: a Project with the org chart, internal glossary, tool guides, key processes, day 1 / week 1 / month 1 checklists and top 30 questions. "Who do I ask for X" routing. A log of common questions so managers see confusion points. Access in the team's chat tool. Monthly refreshes.
Pitch: "Hi [Name], I build onboarding assistants that answer new hires' first-month questions from your own docs, so your team stops answering 'where do I find...' for the 50th time. Want a 5-minute Loom?"
Pitfalls: keep a talk-to-a-human path. Update monthly. Compensation, conflict and performance questions always escalate.

## 8. Meeting brief automation for executives
Suggested price: $8-15K setup, $2-4K/mo per exec, 2 weeks.
Who pays: CEOs, CROs, CMOs and COOs at mid-size companies, founders in back-to-back meetings, executive assistants buying on their behalf.
Deliver: a scheduled workflow that preps a brief the night before (or before each meeting), formatted the way that exec likes. Calendar, email and web search connectors. Delivery by email, Slack or a doc. An optional end-of-day recap. EA-friendly editing. A quarterly review.
Pitch: "Hi [Name], I build automated pre-meeting briefs for execs: who you're meeting, what they care about and the one question to ask, waiting the night before. Want a sample?"
Pitfalls: never let it invent facts about people. Link sources. QA every brief for the first 30 days. Skip routine internal meetings.

## 9. Personal AI setup for one executive
Suggested price: $5-10K build, $500-1.5K/mo maintenance, about 1 week.
Who pays: VPs and up at 100+ person companies, founders and CEOs, partners at consulting, law and agency firms.
Deliver: a 90-minute discovery interview. 20+ samples of their writing for voice. A Project with their priorities, KPIs, strategy, key relationships and modes for their most common tasks (board memos, 1:1 prep, report summaries). Gmail, Calendar and Drive connectors. A 2-hour onboarding for the exec and their EA. 30 days of tuning, then monthly refreshes.
Pitch: "Hi [Name], I build personal AI assistants for executives, trained on your voice and priorities, so drafts, prep and summaries come out the way you'd write them. Want a 5-minute look at one?"
Pitfalls: the voice match is everything. Be honest that it speeds up the work around their judgment. Sensitive topics (board, comp, HR) always go back to them.

## 10. Internal knowledge base with AI search
Suggested price: $10-20K setup, $2-5K/mo retainer, 3-4 weeks.
Who pays: COOs and heads of operations at 50-300 person companies, founder-led companies where knowledge lives in one head, remote teams.
Deliver: an audit of existing docs (Notion, Confluence, Drive, Slack). A keep / kill / merge / rewrite plan. A Project or connected workspace with the cleaned docs. Access in the team's chat tool. Routing to the right person when docs don't cover it. A monthly report of unanswered questions. Team training. Monthly maintenance.
Pitch: "Hi [Name], I set up AI search over your company's own docs, so people stop asking 'where do I find X.' Want a 5-minute Loom?"
Pitfalls: clean the docs first. It must say "that isn't documented" instead of guessing. Get IT and security on board early for any company over 100 people.

---

## How to stack services
- HR stack: 1 (HR assistant) then 7 (onboarding) then 3 (training). Same buyer.
- Sales stack: 2 (email drafting) then 4 (prospect briefs) then 8 (briefs for the sales VP).
- Exec stack: 9 (personal setup) then 8 (meeting briefs), then the rest of the C-suite.
- Company-wide stack: 10 (knowledge base) then 7 (onboarding) then 1 (HR). One audit opens three projects.